TurinTech recently collaborated with Caplin, a cutting-edge trading solutions provider, to run an AI hot desk for a Caplin Hackday. Caplin hackdays give some time and opportunities to the engineers to showcase their skills and creativity by exploring innovative ideas that can even make their way into final products.
The two teams worked together for 24 hours, and within that time span, managed to trial a few key machine learning-based solutions to AI enable some of Caplin’s current sales/trader workflows. Caplin team primarily consisted of engineers, and it was great to see the team use evoML with very little background in machine learning.
“We realised that with evoML we can easily and quickly AI-enable our workflows and prototype new ideas. Without evoML, what we delivered within 24 hours probably would have taken weeks or months.”
– Stephen South, CTO, Caplin
In a typical data science pipeline, building and evaluating a machine learning model can take weeks or even months. However, with evoML, Caplin teams were able to easily build, expose and test machine learning models to prototype how advanced AI-based analytics can support and help their client’s trading decisions.
Here we summarise two Caplin projects that used evoML to incorporate advanced machine learning-based insights into their trading platform:
In the current financial and trading ecosystem, AI-based analytics can bring immense value-add to businesses. However, using conventional data science approaches to develop AI-based solutions can be extremely time consuming and costly, even if it just to explore the feasibility of incorporating AI into a pipeline. Business R&D teams should ideally explore ways to trial at speed; and when things do not go as planned, the best way forward is to try and fail cheap and fail fast.
This hackathon gave a valuable opportunity to the teams to quickly experiment and prototype machine learning-based approaches to improve trading decision-making. The hackathon was evidence that evoML’s automated and efficient machine learning model building tools are ideal for companies that want to avoid time consuming and cumbersome approaches to trialling AI.
With evoML, even teams without a strong background in data science can easily and quickly trial new AI features for their products. This enables businesses to empower more teams to easily innovate leveraging their domain knowledge. The machine learning models built with evoML come with production-ready model code written in Python, enabling teams to easily implement the solution at production level. This reduces the strain on software engineers on integrating machine learning solutions within their existing pipelines.
